This proposal, submitted under the auspices of the program supporting Small Grants for Pilot Research, seeks funds to develop, refine and validate analytic techniques for inferring the relative contributions of feedforward and feedback contributions to fMRI activations in multiple visual areas. Specifically, we will use nonlinear equations to model activations (blood oxygen-level dependent signals registered using functional magnetic resonance imaging) measured while human observers report fluctuations in visual perception associated with viewing binocular rivalry (i.e., the alternations in visual awareness between two dissimilar images viewed by the separate eyes). These analytic techniques have been described in the literature (Buchel & Friston, 1995) but have been rarely employed in studies of vision, despite their potential for significantly extending the range of conclusions drawn from brain imaging studies. Given the dramatic increase in the use of fMRI for the study of vision and its disorders, it is imperative to develop, refine and validate techniques that extend the range of problems that can be addressed using fMRI. The present proposal has this potential.